In multidimensional geriatric assessment, elderly patients are evaluated from different points of view to decide whether to accept or reject their requests for special benefits. The decision is taken in several stages and it does not currently rely on any decision support system. The need to make it more equitable and consistent gave rise to this case study, based on data relating to a retrospective cohort of patients over 67 years of age, who underwent a geriatric examination and geriatric multi-dimensional assessment. To avoid selection and evaluation bias, it was deemed appropriate to consider only patients evaluated with the same methodology and measurement scales. In a preliminary step some inconsistencies were identified in the available data-set. Then, a subset of patients in need for a re-evaluation was defined by a suitable combination of different methods: namely, threshold values related to single indicators, dominance analysis and search for inconsistent pairs. This subset of patients was re-evaluated and a tentative classification model was developed. Therefore, the outcome of the case study was a critical review of the data, the model and the decision-making process.